Raman spectroscopy of ZnS quantum dots

Raman spectroscopy of ZnS quantum dots

Journal of Alloys and Compounds 637 (2015) 401–406 Contents lists available at ScienceDirect Journal of Alloys and Compounds journal homepage: www.e...

928KB Sizes 3 Downloads 72 Views

Journal of Alloys and Compounds 637 (2015) 401–406

Contents lists available at ScienceDirect

Journal of Alloys and Compounds journal homepage: www.elsevier.com/locate/jalcom

Raman spectroscopy of ZnS quantum dots J. Trajic´ a,⇑, R. Kostic´ a, N. Romcˇevic´ a, M. Romcˇevic´ a, M. Mitric´ b, V. Lazovic´ a, P. Balazˇ c, D. Stojanovic´ a a

Institute of Physics, University of Belgrade, 11080 Belgrade, Serbia Institute Vincˇa, University of Belgrade, 11000 Belgrade, Serbia c Institute of Geotechnics, Slovak Academy of Sciences, 043 53 Košice, Slovakia b

a r t i c l e

i n f o

Article history: Received 2 February 2015 Received in revised form 2 March 2015 Accepted 3 March 2015 Available online 10 March 2015 Keywords: Nanostructured materials Scanning electron microscopy X-ray diffraction Light absorption and reflection

a b s t r a c t ZnS nanoparticles were synthesized mechanochemically by high-energy milling. In order to investigate influence off milling time to sample properties, samples were produced in three different milling times (5 min, 10 min and 20 min). The morphology of samples has been investigated by scanning electron microscopy (SEM) and high resolution transmission electron microscopy (HRTEM). X-ray diffraction (XRD) investigation of synthesized nanocrystals identified cubic structure. From XRD, ZnS size of crystallites was estimated as 1.9 nm (after 5 min milling time), 2.3 nm (10 min) and 2.4 nm (20 min), implying that we are in strong confinement regime. The optical properties were studied by Raman spectroscopy, in spectral region 100–500 cm1, excitation source was 514.5 nm (EL = 2.41 eV), which means that we are in off resonant regime. Dominant spectral structures, of comparable intensity, are registered in spectral region 130–180 cm1, around 265 cm1 and around 345 cm1. First two are assigned as second-order ZnS modes. A theoretical model of continuum medium was used to calculate frequencies of the confined optical phonons in ZnS. Satisfactory agreement with experimental results was found and mode at 345 cm1 is assigned as LO type phonon confined in ZnS nanocrystal. Ó 2015 Elsevier B.V. All rights reserved.

1. Introduction The preparation and characterization of different chalcogenides have attracted considerable attention due to their important unique physical and chemical properties [1–4]. Research on semiconductor nanoparticles stimulated great interest in recent years because of their unique optical and electrical properties. Among the semiconductor nanoparticles, zinc sulfide (ZnS) as an important II–VI semiconductor has been investigated extensively because of its broad spectrum of potential applications such as in catalysts, cathode-ray tubes (CRT) and field emission display (FED) phosphors for a long time. It can also be used for electroluminescent devices and photodiodes [5–7]. Nanoparticles differ from bulk particles because of the high surface to volume ratio, which induces the structural and electronic changes. These differences depend on particle sizes, shape and surface characteristics. The decrease of particle sizes causes an extremely high surface area to volume ratio. The enhanced surface area increases surface states, which change the activity of electrons and holes, and affects the chemical reaction dynamics. Therefore, much research on ZnS particles and their physicochemical

⇑ Corresponding author. Tel.: +381 11 3713035; fax: +381 11 3713 052. E-mail address: [email protected] (J. Trajic´). http://dx.doi.org/10.1016/j.jallcom.2015.03.027 0925-8388/Ó 2015 Elsevier B.V. All rights reserved.

properties has been carried out and various methods have been used for the preparation of these nanoparticles [8–13]. In resonant Raman spectroscopy energy of the incident laser light is close to energy of electronic transition. Energy of basic interband transition (1sh–1se) in the quantum dots increases as the dimension of the dot decreases. Simple model based on effective mass approximation can estimate transition energy of QD [14]. This energy depends on the dimension of the dot, and parameters of the bulk material like energy gap, dielectric permittivity, and electron and hole effective masses. If one uses parameters characteristic for bulk ZnS (Eg = 3.66 eV, e = 8.1, meeff = 0.28 and mheff = 0.49) [15,16], QD transition energy E(1sh–1se) for dimension 2 nm is over 4 eV. If energy of the incident laser light is smaller than energy of electronic transition, as in this case, Raman spectroscopy is in off resonance regime. Exciton Bohr radius of material is the measure of confinement. If a size of QD is smaller than exciton Bohr radius the dot is in a strong confinement regime and energy spectrum is discrete. ZnS exciton Bohr radius, for parameters given in previous paragraph, is 2.5 nm. So we are investigating ZnS QDs in strong confinement regime. In this paper we report Raman spectroscopy studies of the ZnS nanoparticles which are mechanochemical synthesized using high-energy milling. Samples characterization was performed using X-ray diffraction (XRD), scanning electron microscopy

J. Trajic´ et al. / Journal of Alloys and Compounds 637 (2015) 401–406

402

(SEM) and high resolution transmission electron microscopy (HRTEM), while optical properties were analyzed using Raman spectra measurements.

magnification. In the case in Fig. 2(b) a square contrast is found in the center of the micrograph with interplanar distances of 0.27 and 0.28 nm, which implies a region with axis zone near to direction [0 0 1] [17].

2. Samples preparation and characterization 2.3. X-ray diffraction (XRD) Mechanochemical synthesis of ZnS nanoparticles was performed in a Pulverisette 6 planetary mill. The milling condition were: 50 balls of 10 mm diameter; weight charge of total powder mixture in the mill was 14.2 g, ball charge in the mill was 360 g, material of milling chamber and balls was tungsten carbide and rotation speed of the planet carrier was 500 rpm. Milling time was 5, 10 and 20 min using an argon atmosphere as a protective medium in the mill [17]. 2.1. Scanning electron microscopy (SEM) The morphology of samples has been investigated by SEM using high resolution electron microscope MIRA3 FEG-SEM, Tescan at accelerating voltage lower than 29 kV. Before that, the surface of samples was coated with an ultrathin gold layer using SC7620 Mini Sputter Coater, Quorum Technologies, with the purpose to prevent the accumulation of static electric fields at the specimen due to the electron irradiation required during imaging. Micrographs of ZnS nanoparticles observed by SEM are presented in Fig. 1. As it can see powder is composed by well-defined and separated nanoparticles. The clusters and nanoparticles are clearly visible. These nanoparticles are spherical and have about 2 nm of diameter, which is close to the microscope resolution limit. Increase in milling time causes better nanoparticles separation, but their dimension remains almost unchanged. 2.2. High resolution transmission electron microscopy (HRTEM) High resolution TEM (HRTEM-Philips Tecnai 200 operated at 200 kV), is an excellent method to study metal sulfide semiconductor nanostructures, where core–shell or stoichiometric systems can be distinguished [18,19]. HRTEM images determine the size of the nanoparticles [20], the type of structures produced [21], and also the morphologies that are possibly induced [22]. The samples were not covered by any type of conductive material to maintain their original properties. In Fig. 2 two different micrographs are shown. The HRTEM images for all three samples are very similar and we choose to present only one. In Fig. 2(a) an area of 16 nm  16 nm is observed. Several clusters are clearly identified, and particularly three of them having sizes of 2.6, 3.7 and 3.4 nm respectively. The corresponding fast Fourier transform (FFT) denotes the polycrystalline material, which must be composed of nanocrystals. Lattice distance of the samples can be determined by applying a higher

(a) 5 min

(b) 10 min

The structural characteristics were obtained by the XRD powder technique. All samples were examined under the same conditions, using a Philips PW 1050 diffractometer equipped with a PW 1730 generator, 40 kV  20 mA, using Ni filtered Co Ka radiation of 0.1778897 nm at room temperature. Measurements were carried out in the 2h range of 10–100° with a scanning step of 0.05° and 10 s scanning time per step. The X-ray diffraction patterns of the ZnS powders obtained after various milling times are presented in Fig. 3. Diffraction patterns show mainly the reflection of cubic phase, according to card JCPDS 03-0524. The refracting planes denoted with (h k l) indices are 1 1 1, 2 2 0 and 3 1 1, respectively. Some divergence from the compared results can be explained by the fact that X-ray powder diffraction analysis gives a statistical result and that samples are with smaller size than as usually. Using the X-ray Line Profile Fitting Program (XFIT) with a Fundamental Parameters convolution approach to generating line profiles [23] the coherent domain sizes of the synthesized powders were calculated. ZnS crystallite size was estimated to 1.9 nm (after 5 min milling time), 2.3 nm (10 min) and 2.4 nm (20 min). Nanocrystallite sizes estimated from XRD spectra are in agreement with values obtained by applied microscopic methods. 3. Results and discussion The Raman spectra of ZnS powders obtained in a manner described in previous chapter were measured in the spectral range 100–500 cm1 at room temperature. The micro-Raman spectra were taken in the backscattering configuration and analyzed by Jobin Yvon T64000 spectrometer, equipped with nitrogen cooled charge-coupled-device detector. As an excitation source we used the 514.5 nm (2.41 eV) line of an Ar-iron laser. This excitation energy is in off-resonance regime even in bulk ZnS. It is clear, that in the QD case Raman spectroscopy is very far from the resonant regime. As we expected, registered spectral features were of low intensity. The measurements were performed at different laser power in order to optimize the signal in the whole spectral region 100–500 cm1. The Raman spectra of ZnS powders obtained after various milling times, in the spectral range from 100 cm1 to 500 cm1, are presented in Fig. 4. Experimental Raman scattering spectra are analyzed by the deconvolution to Lorentzian curves [24].

(c) 20 min

Fig. 1. SEM images of ZnS nanoparticles obtained after milling time of 5 min (a), 10 min (b) and 20 min (c).

J. Trajic´ et al. / Journal of Alloys and Compounds 637 (2015) 401–406

403

Fig. 2. HRTEM analysis of mechanochemically synthesized ZnS nanoparticles: (a) identification of nanoparticle with size around 3 nm and (b) determination of structure using the interplanar distance measurement.

(111)

(220) (311) Intensity [arb. un.]

ZnS, 5 min. d=1.9 nm

ZnS, 10 min. d=2.3 nm

group of the cubic unit cell is F43mðT 2d Þ. The primitive unit cell is trigonal and contains only one formula unit i.e. two atoms. That is why the structure has six degrees of freedom, three acoustic and three optical. In the first order Raman effect only phonon wave vectors very near the Brillouin zone center (BZC) can participate. For zinc-blende structure at the BZC (point C), both acoustic and optical modes are triply degenerate, and have symmetry species C15(F2). In polar crystals, like ZnS, the macroscopic electric field associated with LO vibrations makes the LO mode energy greater than the TO mode energy. This effect removes triply degeneration in the BZC, producing doubly degeneration of TO mode and single degeneration of LO mode. The optical modes which we expect to see in the first-order Raman scattering of bulk sample are double degenerate TO and a nondegenerate LO phonon. Frequencies of these modes in ZnS are well established through calculations [25,26] and experimentally by polarized Raman [26–28] or

0

20

40

60

80

345

265

ZnS, 20 min. d=2.4 nm

5 min 175

100

2θ [deg.] 310

Black thick line presents resulting spectral curve. Positions of Lorentzians are given above the curves in Fig. 4. Dominant wide structures in experimental spectra are: wide multimodal feature in region 130–180 cm1, wide structure centered at 265 cm1 and wide structure centered at 345 cm1, Fig. 4. These ‘‘dominant wide structures’’ in experimental spectra are practically of small intensities. These dominant structures are analyzed in detail. In spectral region 220 cm1 there is a feature in spectra of 10 min and 20 min milling sample. This mode is hardly visible in 5 min milling sample. Also in a region 310 cm1, there is a small intensity feature. These structures are briefly discussed. In order to identify properly experimentally registered vibrational modes we will briefly review vibrational properties of cubic structure bulk ZnS. ZnS crystallizes in cubic (zinc-blende, sphalerite, b-ZnS) or hexagonal (wurtzite) structure. Lattice constant of cubic cell is: a = 5.4 Å. The cubic unit cell contains four formula units. The space

Intensity (arb. units)

Fig. 3. XRD spectra of ZnS powders obtained after various milling times.

220

10 min

multimode

20 min

130 180 100

200

300

400

500

Raman shift (cm-1) Fig. 4. Raman spectra of ZnS powders obtained after various milling times – 5 min: d = 1.9 nm, 10 min: d = 2.3 nm and 20 min: d = 2.4 nm.

J. Trajic´ et al. / Journal of Alloys and Compounds 637 (2015) 401–406

neutron scattering [25] studies: xLO  350 cm1 and 1 xTO 275 cm . If we are in resonant regime Raman scattering of LO mode will be dominant and sometimes TO mode is hard to detect especially in nonpolarized spectra. If we are in non-resonant regime all intensities decrease. Intensity of TO mode becomes even smaller and practically only LO mode is registered. Registered LO mode is wither and also of lower intensity than in resonant case. According to [26], ratio of the TO and LO mode integrated Raman intensities for excitation energy EL = 2.41 eV is about 0.1, as consequence of high value of corresponding electron–phonon deformation potential (DP), obtained after taking into account the antiresonance. In the second order Raman effect momentum conservation involves two phonons. The second order Raman selection rules must be satisfied. The selection rules of the two phonon states at critical points are derived from the reducible direct product representation of the corresponding one phonon states. Scattering process originates from the BZC (C) or from the BZ boundary points of face centered cubic lattice as: point X(D2d), point L(C3v), point W(S4), or critical directions as: direction R(C toward K) [1 1 0]. Double degeneration of transversal modes in C point results in two traversal branches (both optical and acoustical) in direction R. Upper branch is often assigned by index u and the lower branch by index l. As wave vectors at the BZ boundary are much larger than the wave vector of the excitation used in experiments, two phonons created or destroyed in the second order scattering that satisfy momentum conservation, originate from the same point of the Brillouin zone. The combined states density tends to be large at the critical points or critical directions on the BZ boundary. Detailed investigation of ZnS vibration properties, that include calculation (based on the bond charge model lattice dynamics) of the densities of one- and two-phonon states and polarized Raman scattering measurements, is presented in [26]. Raman spectra of bulk ZnS samples with different isotopic compositions, their dependence on temperature and pressure and enable reliable assignation of observed Raman features. Strong DP for two-phonon process of cubic ZnS (two orders of magnitude higher than in the single-photon process) is responsible for the increase in the twophonon scattering. These information about structure are vibrational properties of bulk ZnS are the starting points for analysis that concerns ZnS QDs. In nanocrystals optical modes are confined, bulk selection rules are ruined, high surface to volume ratio increase the role of surface properties, but there is fundamental track of bulk properties. Analysis of the Raman spectra presented in Fig. 4 starts from the optical phonon region i.e. region 275 cm1 (xTO) to 350 cm1 (xLO). A continuum model of the optical phonon confinement in QD is used. Parameters were transferred from the bulk phonon dispersion curves. It is limited to nanoparticles of regular shape. Although this is not the case in real nano-crystallites, we present results of calculation for ideal spherical ZnS QD. One small spherical ZnS crystal, isotropic and homogeneous inside, is considered. This consideration of confined optical vibrations in nanocrystals is based on macroscopic equation for the relative displacement of the positive and negative ions [29,30]. Parameters of this macroscopic equation are: reduced mass density, xTO: the TO bulk frequency, the transverse charge, the unit cell volume, and bT and bL phenomenological bending parameters of TO and LO bulk dispersion curves. This equation is solved in spherical coordinates. The spherically symmetric solutions of equation must belong to the irreducible representations of the three-dimensional rotation-inversion group O(3) labeled as Dgl (even) and Dul (odd upon inversion). The mixed modes belong to D0g, D1u, D2g, . . .. The dipole operator responsible for FIR absorption belongs to D1u while Raman transition operator for allowed

scattering belongs to D0g and D2g [31]. Frequencies of the spherical (l = 0) and spheroidal quadropolar modes (l = 2) can be calculated and in principal observed by resonant Raman scattering. If we assume, as in [32,33], that at the surface of the sphere all components of displacement are almost zero, the electrostatic potential and the normal component of the electric displacement are continuous. After applying these assumptions one can obtain frequencies of the Raman active (l = 0 and l = 2, n = 1, 2, 3, . . .) and FIR-active (l = 1, n = 1, 2, 3, . . .) modes, l and n being the spherical quantum numbers. The most important contribution to one-phonon Raman scattering corresponds to l = 0 (the quadrupole modes are active only under resonance conditions and their contribution is much smaller). This mode is excited for parallel polarizations of the incident and scattered light. The corresponding frequencies are:

x2n ¼ x2LO  b2L

 2 2l n d

ð1Þ

xLO is the LO bulk frequency (xLO = 350 cm1 in ZnS), d is the diameter of the sphere, ln is the n-th node of the Bessel spherical function j1 (l1 < l2 < l3 < . . .). Frequency shift (difference between xn and xLO) for fixed d depends on bL. bL = 2.6 103 m/s for bulk ZnS. xn increases as the dimension of the dot (d) increases, and in the limit: d ? 1 frequencies xn converge to xLO. Fig. 5 presents dependence of optical vibration modes frequencies (l = 0, n = 1, 2, 3) on the diameter of ZnS QD. The smaller the diameter the lower is the frequency of confined mode. As concerns intensity, this model predict the most intensive peak in QD Raman spectra to be the mode x1(n = 1). Exact positions, from the deconvolution of experimental spectra, of mode at 345 cm1 for three dimensions: 1.9 nm (5 min milling time), 2.3 nm (10 min) and 2.4 nm (20 min), Fig. 4, are marked with stars in Fig. 5. It is evident that experimental values are in very good agreement with calculated values. In reality there is QD size distribution, QD shape irregularity, inhomogeneity inside, some interaction between nanoparticles, etc. The more the QD behaves as rigid sphere the vibration modes produce less electric field outside the sphere, there is less interaction associated with this mode between nano-crystals. So, the Raman cross-section of an array of scatters is simply a superposition of their individual contributions, and it is proportional

ωLO

350

340

l=0 ωn [cm-1]

404

330

n=1 n=2 n=3

320

310

ZnS 300 1

2

3

4

5

6

d [nm] Fig. 5. The radial dependence of l = 0 optical modes for ZnS spherical QD.

J. Trajic´ et al. / Journal of Alloys and Compounds 637 (2015) 401–406

to the volume fraction of semiconductor. During milling dimension of ZnS QDs slightly increases, agglomerates and clusters become bigger but more separated, Fig. 1. Volume fraction does not change too much. That is why intensity of mode at 345 cm1 does not change significantly in all three samples, Fig. 4. While analyzing Raman spectra of the sample of the smallest diameter QD (d = 1.9 nm) we find out that adding spectral structure in region 300–330 cm1 (centered at 310 cm1) reproduces experimental spectra better than without it (dashed line in Fig. 5). Similar procedure was applied for the other two samples. We tried to explain origin of this structure. We focused to the spectra of the sample of the smallest diameter QD (d = 1.9 nm). Calculated optical vibration modes of d = 1.9 nm ZnS QD, Fig. 5, are x2 = 330 cm1 (n = 2) and x3 = 340 cm1 (n = 3). Predicted Raman intensity of these modes is much smaller than intensity at 345 cm1 (n = 1) mode. One of possible explanation is that the contribution of these two modes is registered. If we go back to vibrational properties of bulk ZnS we can find out that the frequencies of second order ZnS Raman peaks measured at 304 cm1 and 312 cm1 are assigned as [LA + TAl]W and [LA + TAu]W,R [26]. If non polar matrix surrounds QD, there is one surface mode for each l. In case of ZnS QD without matrix i.e. in vacuum (ematrix = 1) frequencies of l = 1 and l = 2 surface phonons are 330 cm1 and 334 cm1. If assume that dots inside agglomerate are in non polar matrix of dielectric permittivity higher than vacuum, frequencies of surface modes will be lower. Raman scattering spectra of ZnS QD, mean size 2.8 nm, are presented in 29]. The Raman spectra of the structures with ZnS QDs contain single line at frequency 320 cm1. Calculated surface mode frequency (l = 1, ematrix = 2.4 in their case) is 316 cm1, and this mode was identified as surface mode [34]. These are possible origins of this wide, low intensity structure in spectral region 300–330 cm-1. We continue analysis of the Raman spectra presented in Fig. 4 in region wide and relatively strong spectral feature centered at 265 cm1. Spectral structure centered at 265 cm1 is of the same order of intensity as the mode at 345 cm1. Frequencies of bulk second order ZnS Raman peaks established by calculations at 244 cm1, 256 cm1 and 257 cm1 are assigned as [2TAu]R, [LA + TAl]W and [LO  TAu]X [26]. We believe that this wide structure can be a sum of these contributions. Similar spectral structure in bulk ZnS unpolarized Raman spectra was also registered, and assigned as combination mode in point W [28]. Raman scattering studies of ZnS nanoclusters of typical sizes 2.2–5 nm are presented in [35]. Raman spectra with the visible excitation (532 nm) of nanoclusters of the size (d  2.5 nm) i.e. smaller than exciton Bohr radius, shows a structure centered at 256 cm1 comparable in intensity to the structure centered at 350 cm1. For ZnS nanoparticles of d = 2.2 nm structure centered at 256 cm1 dominates over the structure centered at 350 cm1 [35]. There is additional weak spectral feature at 220 cm1, Fig. 4. In case of d = 1.9 nm (5 min milling) this feature is almost invisible. But for larger nanocrystals (d = 2.3 nm and d = 2.4 nm) it is clearly seen. Frequency of second order ZnS Raman peak calculated and registered at 218 cm1 is assigned as [TO  TA]X [26]. The weak but wide feature at about 218 cm1 was registered in Raman spectra of ZnS nanoparticles synthesized using Langmuir–Blodgett technique and attributed to second order scattering [36]. We assigned this peak as second order [TO  TA]X scattering. There is lack of results in spectral region below 200 cm1. It is very difficult to get spectra in this region. Presented spectra were performed at very low laser power. We registered wide feature of irregular shape in region 130–180 cm1. Frequencies of second order ZnS Raman peaks calculated at 137 cm1, 143 cm1,

405

167 cm1 and 180 cm1 are assigned as: 2TAL, [TOu  LA]R, [LO  LA]R and 2TAX [26]. We believe that this wide structure can be a sum of these contributions. Registered integral intensity of this spectral structure, which originate from few second order modes, is comparable to the intensity of confined LO mode. In sample produced after 5 min milling time (d = 1.9 nm) a single weak spectral feature at 175 cm1 is more prominent than in the other samples, Fig. 4. Raman activity in low frequency region, below 200 cm1, is hard to detect, and there are no results of the other groups about QD Raman spectroscopy to compare with. So, feature of irregular shape in region 130–180 cm1 is recognized as group of second order modes. 4. Conclusion We report Raman spectra of mechanochemically synthesized ZnS nanocrystals. Milling time was varied. Dimension of nanocrystals are of 1.9–2.4 nm depending of the duration of milling. Small dimension of ZnS QD results in the strong confinement regime. Raman spectra were measured in off resonance regime. A continuum model of the optical phonon confinement in QD is used for investigation in optical phonon region i.e. region 275 cm1 (xTO) to 350 cm1 (xLO). Despite the fact that this model treats an ideal case, measured frequency of mode at 345 cm1 is in a very good agreement to predicted values. This mode, few cm1 below bulk xLO, is identified as a confined LO (l = 0) mode. As we expect, this mode is of much lower intensity, compared to the bulk ZnS. Registered intensities of multimodal spectral features in spectral region 130–180 cm1 and 265 cm1, are comparable to the intensity of confined LO mode. These multimodal structures are sums of two-phonon Raman scattering from the BZ boundary. ZnS is a system of large deformation potential in the two-phonon processes. That is why there is remarkable Raman activity in spectral region 130–180 cm1 and 265 cm1 of two-phonon scattering processes in ZnS QD, even in off resonant regime. Acknowledgements This work in Serbia was supported by Serbian Ministry of Education, Science and Technological Development under Project III45003. This work was also supported by Slovak Grant Agency VEGA (Project 2/0027/14). References [1] K.J. Klabunde, Nanoscale Materials in Chemistry, Wiley Interscience, New York, 2001. [2] M. Fayette, R.D. Robinson, J. Mater. Chem. A 2 (2014) 5965. [3] X.R. Rui, H.T. Tan, Q.G. Yan, Nanoscale 69 (2014) 9889. [4] J. Chang, E.R. Waclawik, RSC Adv. 4 (2014) 23505. [5] H. Wang, X. Lu, Y. Zhao, C. Wang, Mater. Lett. 60 (2006) 2480. [6] N. Habubi, M. Hashim, A. Al-Yasiri, Baghdad Sci. J. 7 (2010) 1421. [7] J.S. Jie, W.J. Zhang, I. Bello, Ch.-S. Lee, S.T. Lee, Nano Today 5 (2010) 313. [8] P. Balaz, M. Balintova, Z. Bastl, J. Briancin, V. Sepelak, Solid State Ionics 101 (1997) 45. [9] S.A. Chen, W.M. Liu, Langmuir 15 (1999) 8100. [10] N.A. Dhas, A. Zaban, A. Gedanken, Chem. Mater. 11 (1999) 806. [11] L.P. Wang, G.Y. Hong, Mater. Res. Bull. 35 (2000) 695. [12] N.R. Pawaskar, S.D. Sathaye, M. Bhadbhade, K.R. Patil, Mater. Res. Bull. 37 (2002) 1539. [13] J. Chen, Y. Li, Y. Wang, J. Yun, D. Cao, Mater. Res. Bull. 39 (2004) 185. [14] L.E. Brus, J. Chem. Phys. 79 (11) (1984) 5566. [15] M. Sahin, S. Nizamoglu, A.E. Kavruk, H.V. Demir, J. Appl. Phys. 106 (2009) 043704. [16] E.C. Nicolescu, M. Cristea, A. Spandonide, Superlattices Microstruct. 62 (2013) 1. [17] E. Dutkova, P. Balaz, P. Pourghahramani, S. Velumani, J.A. Ascencio, N.G. Kostova, J. Nanosci. Nanotechnol. 9 (2009) 6600. [18] F. Paraguay-Delgado, W. Antunez-Flores, M. Miki-Yoshida, A. AguilarElguezabal, P. Santiago, J.R. Díaz, J.A. Ascencio, Nanotechnology 16 (2005) 688. [19] O. Šolcová, Ch.D. Uecker, U. Steinike, K. Jirátová, Appl. Catal. A 94 (1994) 153. [20] M. José-Yacamán, J.A. Ascencio, H. Liu, J. Vac. Sci. Technol. B19 (2001) 1091.

406

J. Trajic´ et al. / Journal of Alloys and Compounds 637 (2015) 401–406

[21] J.A. Ascencio, C. Gutiérrez-Wing, M.E. Espinosa-Pesqueira, M. Marín, S. Tehuacanero, C. Zorrilla, M. José-Yacamán, Surf. Sci. 396 (1998) 349. [22] J.S. Ascencio, H.B. Liu, U. Pal, A. Medina, Z.L. Wang, Microsc. Res. Tech. 69 (2006) 522. [23] R.W. Cheary, A.A. Coelho, J. Appl. Cryst. 25 (1992) 109. [24] B.H. Henry, J.R. During (Eds.), Raman Spectroscopy: Sixty Years On, vol. 10, Elsevier, Amsterdam, 1990. [25] B.D. Rajput, D.A. Browne, Phys. Rev. B 53 (1996) 9052. [26] J. Serrano, A. Cantarero, M. Cardona, N. Garro, R. Luck, R.E. Tallman, T.M. Ritter, B.A. Weinstein, Phys. Rev. B 69 (2004) 014301. [27] W.G. Nilsen, Phys. Rev. 182 (3) (1969) 838. [28] Y.C. Cheng, C.Q. Jin, F. Gao, X.L. Wu, W. Zhong, S.H. Li, Paul K. Chu, J. Appl. Phys. 106 (2009) 123505.

[29] [30] [31] [32] [33] [34] [35] [36]

M.P. Chamberlain, C. Trallero-Giner, M. Cardona, Phys. Rev. B 51 (1995) 1680. R. Roca, C. Trallero-Giner, M. Cardona, Phys. Rev. B 49 (1994) 13074. E. Duval, Phys. Rev. B 46 (1992) 5795. C. Trallero-Giner, A. Debernardi, M. Cardona, E. Menendez-Proupin, A.I. Ekimov, Phys. Rev. B 57 (1998) 4664. N. Romcˇevic´, R. Kostic´, M. Romcˇevic´, M.I. Comor, J.M. Nedeljkovic´, J. Phys. D: Appl. Phys. 38 (2005) 4321. G. Milekhin, L.L. Sveshnikova, S.M. Repinskii, A.K. Gutakovskii, M. Friedrich, D.R.T. Zahn, Phys. Solid State 44 (10) (2002) 1976. S. Dhara, A.K. Arora, Santanu Bera, Jay Ghatak, J. Raman Spectrosc. 41 (2010) 1102. G. Milekhin, N.A. Yeryukov, L.L. Sveshnikova, T.A. Duda, C. Himcinschi, E.I. Zenkevich, D.R.T. Zahn, Appl. Phys. A 107 (2012) 275.